Piecewise linear approximations of set-valued maps (Q1122076)
From MaRDI portal
| This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use this page instead for the normal view: Piecewise linear approximations of set-valued maps |
scientific article; zbMATH DE number 4105535
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | Piecewise linear approximations of set-valued maps |
scientific article; zbMATH DE number 4105535 |
Statements
Piecewise linear approximations of set-valued maps (English)
0 references
1989
0 references
Let F be a set-valued function, defined on [0,1] with values F(t), which are compact subsets of the Euclidean (real) space \(R^ n\). The question of the approximation of such a map F by one which is piecewise linear is studied. If all the values F(t) are convex sets then the standard linear interpolation formula yields a good approximation. Otherwise the standard interpolation may fail to supply an approximation. To overcome the difficulty the author approaches the problem otherwise. The ensemble of compact sets can be regarded as a metric space with the Hausdorff distance h(A,B) between the two compact subsets A and B in \(R^ n\). The following two problems are studied. (I). Given a partition \(\pi =\{0=t_ 0<t_ 1<...<t_ k=1\}\) find a continuous set-valued function \(F_{\pi}\) such that \(F_{\pi}(t_ j)=F(t_ j)\) for all j, and whenever \(t_ j\leq s_ 1<s_ 2<s_ 3\leq t_{j+1}\), the value \(F_{\pi}(s_ 2)\) is a weighted average of \(F_{\pi}(s_ 1)\) and \(F_{\pi}(s_ 3)\), with respective weights \((s_ 3-s_ 2)/(s_ 3- s_ 1)\) and \((s_ 2-s_ 1)/(s_ 3-s_ 1)\). (II). Given a partition \(\pi\), find a continuous set-valued function \(F_{\pi}\), such that \(F_{\pi}(t_ j)=F(t_ j)\) for all j, and whenever \(t_ j\leq s_ 1<s_ 2\leq t_{j+1}\) then \[ h(F_{\pi}(s_ 1),F_{\pi}(s_ 2))=((s_ 2-s_ 1)/(t_{j+1}-t_ j))h(F(t_ j),F(t_{j+1})). \]
0 references
set-valued map
0 references
piecewise linear approximation
0 references
Hausdorff distance
0 references
weighted average
0 references